Back from a busy couple of days at TechFest 2019 I’m pleased to say it wasn’t just the sea air putting me in a good mood. UKOUG put on another excellent event and I’m delighted to be sharing my review with everyone reading this blog.

Although I’m not usually thinking about such matters at 9AM (Honest), I started my day with ‘Picking a Good Wine for <$20 Using Oracle Autonomous, Machine Learning, and Analytics Cloud’ by Oracle’s Charlie Berger and Rittman Mead’s Francesco Tisiot. This excellent talk was built around the practical application of Oracle’s technology, and I would recommend checking out the algorithms in Oracle’s Autonomous Database, which is free to try, for yourselves. Charlie also informed us that in the near future we’ll be able to enjoy support for languages beyond SQL, including K and Python.

From here I went to Monday’s keynote, preceded by the UKOUG’s official welcome hosted by President Martin Widlake (UKOUG). Of particular interest to me was the announcement of ‘Business Applications Xchange 2020’ to be held at The Oval on  June 15th and 16th, which I know will be of great interest to all my network. Oracle’s Neil Sholay delivered a fascinating keynote, covering topics from the next industrial revolution through to the nature of disruption and the opportunities and limitations presented by automation. The three key messages I took from Neil’s talk were ‘Embrace the Inevitable’, ‘Automation is a Spectrum’ and ‘Re-calibration isn’t failure’.

Given the emerging importance of Chatbots in today’s market place, I made sure to attend ‘Chatbots Best Practices and Design Patterns’ hosted by Avanttic’s Ruben Rodriguez, David Perez, and Samuel Campos. It was great to see the inner workings of a chatbot, and I was particularly impressed with the ability to identify the language of a user based on their greeting, and using this to select the language spoken by the bot. As explained by Ruben, the key to successful Bot design is to always ask ‘What will your chatbot be used for, and who is going to be using it?’

My last talk before lunch was the Development Community Keynote: Bee Project Update sponsored by Gospel Technology (Andy Clark, Oracle) and I encourage everyone to read up on this extremely important work tackling dwindling bee numbers around the world.

I began the afternoon with Machine Learning – How Data Scientists Improve their Accuracy hosted by Tony Heljula and Ingrid Linros from Peak Indicators. It was extremely helpful to be shown through the entire Machine Learning Lifecycle, and to learn about some of the techniques used in model refinement. Tony was quick to point out the complexity of this topic, and to highlight that even big names around the world can find themselves struggling to produce accurate modelling techniques. One point which I found particularly interesting was the need to account for biases within sampling (such as addressing gender bias amongst job applicants, for example) which may, counter-intuitively, require feeding in rows of data more than once to produce a more accurate model.

Appropriate and efficient use of Data was a key theme throughout the conference, and the next talk I attended on Monday afternoon was ‘Data – The Missing Ingredient’ (Jon Mead, Rittman Mead). The most important messages I took from Jon’s presentation centred around the business use of data – that is to say, the reason for analysing data in the first instance. Jon made the point that this has to be established before seeking source data, as data is ultimately another input into a system and must therefore be controlled to produce a reliable result. Jon also added that ‘whatever your business does with data, it must be accessible to the business user in order to be useful’ – and I’m sure Jon and I aren’t the only two UKOUG members who believes Data Architects are often under-utilised.

I went next to ‘When Machine Learning meets Graphs Databases’ (Gianni Ceresa, Datalysis) who reiterated the point that Machine Learning can only be as good as the data it is fed with. Or, as he termed it ‘Machine Learning is not simply Machine Guessing’.

My last talk on Monday was ‘React Vs Oracle JET for Modern Front End Development’ hosted by Mark Waite and Daniel Curtis from Griffiths Waite. In addition to providing a comprehensive comparison of these technologies, I thought it was particularly valuable that this talk made it clear that choosing the right technology for your business has to go beyond individual preference. Griffiths Waite for example ‘loved JET so much we wrote a book about it’, but they have since abandoned this technology in favour of React because ‘the market has spoken, and done so to such a point that progressing with JET is now no longer worthwhile’.

There was a lively turnout for the networking drinks at the end of the day’s presentations and it was good to meet so many new people whilst also catching up with old friends and associates – although I was very restrained and retired early in anticipation of another packed day on the Tuesday which you can read about in my next blog post.